System for Measuring a Ground Material Health Level

This application is a divisional of U.S. patent application Ser. No. 18/651,466, filed Apr. 30, 2024, which is incorporated herein by reference in its entirety.

Soil, also commonly referred to as earth or dirt, is a mixture of organic matter, minerals, gases, liquids, and organisms that together support the life of plants and soil organisms. Some scientific definitions distinguish dirt from soil by restricting the former term specifically to displaced soil. Soil consists of a solid phase of minerals and organic matter (the soil matrix), as well as a porous phase that holds gases (the soil atmosphere) and water (the soil solution). Accordingly, soil is a three-state system of solids, liquids, and gases. Soil is a product of several factors: the influence of climate, relief (elevation, orientation, and slope of terrain), organisms, and the soil's parent materials (original minerals) interacting over time. It continually undergoes development by way of numerous physical, chemical, and biological processes, which include weathering with associated erosion. Given its complexity and strong internal connectedness, soil ecologists regard soil as an ecosystem.

The technologies described herein will become more apparent to those skilled in the art from studying the Detailed Description in conjunction with the drawings. Embodiments or implementations describing aspects of the invention are illustrated by way of example, and the same references can indicate similar elements. While the drawings depict various implementations for the purpose of illustration, those skilled in the art will recognize that alternative implementations can be employed without departing from the principles of the present technologies. Accordingly, while specific implementations are shown in the drawings, the technology is amenable to various modifications.

The disclosed technology relates to a system for measuring and calculating the health of a ground material. For instance, ground material can include soil, topsoil, subsoil, dirt, earth, clay, sand, or silt. The system uses an auger to measure the ground material. The auger includes a shank, which can be cylindrical and extend vertically into the ground material when taking a measurement. The auger also can include an auger blade attached to the shank. The auger blade can have a twisted shape that wraps or spirals around the shank. The twisted shape of the auger blade helps break up the ground material, allowing for easier measurement of the ground material. The auger can include at least one sensor configured to measure multiple characteristics of the ground material. Knowing the multiple characteristics of the ground material can indicate the health level of the ground material and what plants can grow and be cultivated in the ground material.

For example, the multiple characteristics measured by the sensor can include a nutrient level, a chemical composition, a moisture level, a temperature, a compaction level, and a potential of hydrogen (pH) level of the ground material. The nutrient level can indicate the number of nutrients found in the ground material. The nutrients measured can include nitrogen, phosphorus, or potassium. A higher nutrient level can indicate that the ground material is more suitable for cultivating crops or plants. The chemical composition can indicate the level and type of inorganic or organic matter found in the ground material, which can be used as an indicator of the health of the ground material. The moisture level can indicate the amount of water in the ground material, which can be used to determine the suitability of cultivating a particular plant. The temperature of the ground material can indicate what plants can grow or germinate in the ground material. The compaction level or density of the ground material can indicate a plant's likelihood of penetrating the ground material and being able to grow. The pH level of the ground material can indicate how the elements in the ground material will affect the plant and the plant's ability to grow or germinate. For example, a ground material with a low pH can cause many elements or nutrients to become less available to the plants while causing elements such as iron, aluminum, or manganese to become toxic to the plants.

In one embodiment, the sensor used to measure the multiple characteristics can be a spectrometer. A spectrometer is any instrument used to measure the variation of a physical characteristic over a given range, i.e., a spectrum. In another embodiment, the sensors used to measure the multiple characteristics can include a range sensor, a moisture sensor, a nutrient sensor, a force sensor, a camera, a temperature sensor, a pH sensor, and/or a chemical sensor. The auger can include a sensor window positioned vertically on the auger shank. The sensor window can be configured to receive the ground material. In one embodiment, the sensor window is constructed from a transparent, durable material such as glass, tempered glass, laminated glass, polycarbonate, acrylic, or any other transparent material. In another embodiment, at least one sensor can be located behind the sensor window. In yet another embodiment, the sensor window can include a non-stick coating configured to prevent the ground material and debris from sticking to or blocking the sensor window. The non-stick coating can be clear and attached to the sensor window. In yet another embodiment, the sensor window can include a brush configured to remove ground material and debris from the sensor window. For example, the brush can be attached to a motor to rotate the brush in order to remove the ground material and debris.

The sensor window can be located on the shank of the auger. In one embodiment, the auger can include a blade scoop configured to direct the ground material onto the sensor window. For example, the blade scoop can be constructed from a section of the auger blade. The blade scoop can be a vertically folded section of the auger blade forming an angle between the auger blade and the blade scoop. The blade scoop can be a perpendicular section of the auger blade and act as a channel for the ground material. The blade scoop can help ensure that a proper amount of ground material is channeled past the sensor window to allow the sensor to make an accurate measurement. For example, inaccurate measurements can cause inaccurate data, leading to lower crop yields, slower-growing plants, and/or plant death.

The system can include an auger controller configured to control the auger. The auger controller includes a motor configured to control the speed of the auger. For instance, the motor can control the rotational speed of the auger by setting the revolutions per minute (rpm) at which the auger rotates. The motor can control the translational speed of the auger. For example, translational speed is the speed at which the auger enters the ground material. The compaction level of the ground material can be calculated by monitoring the load experienced by the motor during the drilling process caused by the auger's interaction with the ground material. Additionally, a load sensor or similar sensor can be added to increase the accuracy of the compaction level calculation. The auger controller can include an accelerometer configured to measure the vibration generated by the auger during the drilling processes. For example, different ground material types or health levels can cause the accelerometer to measure different vibration levels.

The auger controller can include a gyroscope. The gyroscope can be configured to control the drilling angle of the auger. For example, the gyroscope can determine the needed drilling angle for the auger and the current angle of the auger. The gyroscope can adjust the angle of the auger to position the auger to the needed drilling angle. The auger controller can include a power manager configured to control the amount of energy consumed by the system. The auger controller can include a central processing unit (CPU) and a graphics processing unit (GPU). For example, the CPU and the GPU can perform the calculations and processes needed for the motor to determine the compaction level, the accelerometer to determine the vibration level, or the gyroscope to determine the drilling angle. In one embodiment, the auger controller can receive the multiple characteristics measured using the auger and locally calculate the ground material's health level using the CPU and GPU.

The system can include a distance sensor configured to detect the distance or depth from the auger to the ground. For example, the distance sensor can be a range sensor, a camera, a radar sensor, a laser distance sensor, an ultrasonic sensor, and/or any sensor capable of measuring a distance. In one embodiment, the distance sensor is externally mounted to the auger controller. In another embodiment, the distance sensor is mounted to a vehicle. The system can include a vehicle configured to transport the auger and auger controller to the location where a health level will be measured. In one embodiment, the vehicle includes any multi-wheeled powered vehicle capable of transporting the auger and auger controller from one location to another. In another embodiment, the vehicle is an existing piece of equipment, such as a tractor, a car, a truck, an all-terrain vehicle (ATV), or any commercially available vehicle capable of driving off-road. For example, the vehicle can transport the auger and auger controller to a location to measure the health of the ground material. The vehicle includes a mounting structure to couple the auger and auger controller to the vehicle. The vehicle can include a global positioning system (GPS). The GPS can navigate the vehicle to a desired location where a health level can be measured. The vehicle can include a wireless modem configured to connect the system to a Wi-Fi, telecommunications, or wireless network. For example, connecting the vehicle to a wireless network can allow a user to control the vehicle or view the vehicle's location remotely. The wireless modem can also allow the vehicle to remotely receive navigation instructions which autonomously control the movement of the vehicle. The vehicle can include multiple vehicle sensors configured to control the vehicle's location autonomously. The multiple sensors can allow the vehicle to navigate to a desired measurement point autonomously.

To measure the soil health at a location, the system receives the GPS coordinates for a geographic area that contains the ground material to be measured. The system can determine a measurement plan that includes at least one measurement point and the location of the measurement point. The system executes the measurement plan, causing the system to navigate to the location of the measurement point. In one embodiment, the system navigates to the location using the vehicle. When at the measurement location, the system can detect the presence of any debris in the auger's drilling path. For example, debris can include any substance separate from the ground material that could damage the auger, such as rocks, metal, or tree roots. When debris is detected, the system can adjust the location of the measurement point. For example, the measurement point can be adjusted to the closest point from the original measurement point that avoids the debris.

The system can cause the auger to enter the ground material, where the sensor can measure the multiple characteristics of the ground material. The system calculates, using, for example, a machine-learning model (ML) model, the ground material's health level based on the ground material's measured characteristics. In one embodiment, the ML model can be trained using ground material control samples. The ground material control samples can include ground material of different types with varying measurements for the multiple characteristics and, therefore, varying health levels.

In one embodiment, the system can generate a gradient map of the ground material's health level for an area's measurement point. The gradient map can indicate the health level at the measurement point for multiple depths below the ground surface of the ground material. In another embodiment, the system can calculate the health level for multiple measurement points at multiple areas. The system can rank the multiple measurement points based on the health level of the ground material at the measurement point. The system can generate a choropleth map of the multiple areas. A choropleth map uses differences in shading, coloring, or the placing of symbols within predefined areas to indicate the average values of a characteristic, property, or quantity in those areas. The choropleth map generated by the system can indicate which sections or parts of the multiple areas have a suitable health level to cultivate a plant.

In yet another embodiment, the system can log the measurements of the multiple characteristics and the health level of the ground material at each of the multiple measurement points for an area. The system can perform a second measurement of the multiple characteristics and the health level of the ground material at the same measurement points in the area. For example, the second measurement can occur after a predetermined time period has occurred, such as one day, one week, one month, one year, or after a planting season. The system can compare the logged measurements of the multiple characteristics and the health level to the second measurements of the multiple characteristics and the health level of the ground material. The system can generate a ground material report indicating a change in the health level of the ground material for the area over the predetermined time period. For example, the ground material report can include how each of the multiple characteristics of the ground material changed over the predetermined time period.

In one embodiment, the system can determine an action plan for the area based on the health level of the soil. The action plan can include performing more measurements in the area, generating a cultivation plan, or generating a ground material rejuvenation plan. For example, the cultivation plan can include a recommendation of different plant species or crops suitable for the area. For example, the ground material rejuvenation plan includes a list of deficiencies in the ground material and a list of techniques to reduce the deficiencies. For example, deficiencies in the ground material can include a lack of vital nutrients, too much of a particular or multiple nutrients, the presence of a contaminant in the ground material, or an improper ground material compaction level. The system can, therefore, reduce an amount of greenhouse gases by increasing plant growth. Plants remove greenhouse gases such as carbon dioxide from the atmosphere. The system allows for the better cultivation of plants and crops by determining the location and type of plant to be cultivated, thereby increasing plant yield and thus decreasing an amount of greenhouse gases.

The description and associated drawings are illustrative examples and are not to be construed as limiting. This disclosure provides certain details for a thorough understanding and enabling description of these examples. One skilled in the relevant technology will understand, however, that the invention can be practiced without many of these details. Likewise, one skilled in the relevant technology will understand that the invention can include well-known structures or features that are not shown or described in detail to avoid unnecessarily obscuring the descriptions of examples.

is an isometric view of a preferred embodimentof the system in which the system is mounted to a vehicle. The system includes auger. Augeris mounted below auger controller. For example, augercan be attached directly to the underside of auger controller. Auger controllercan have a cubic-shaped housing that protects the auger controller and a multitude of sensors. Augerand auger controllercan be mounted to the vehicle using the mounting hardware. Mounting hardwarecan attach to the outer surface of the auger controller housing. For example, mounting hardwarecan include screws, bolts, rivets, or any suitable hardware capable of securing auger controllerto the vehicle.

The vehicle can include wheel mountsand(wheel mounts). Wheel mountscan be configured for large off-road wheels,and(wheels). An ample clearance allows wheel mountsto receive multiple types of wheelsof different sizes, allowing the vehicle to traverse a large variety of terrain. Wheelscan be configured to travel off-road without damaging crops and plants. For example, wheelscan be configured to position the vehicle body above the top point of the crops to prevent damage to the crops while driving. Additionally, wheelscan be spaced appropriately to allow the vehicle to navigate on the same tracks as other farm equipment. For example, a track is a strip of land or ground material where no crops were planted to provide a path for farm equipment.

The vehicle can include a wireless modem. Wireless modemcan provide cellular or wireless connectivity to the vehicle, allowing the system to be controlled remotely. Wireless modemcan be connected to the auger controller and can include GPS capabilities. Wireless modemallows the vehicle to be navigated using a remote control or autonomously without user input. Wireless modemcan allow the system to transmit data over a telecommunications network, Wi-Fi network, or other wireless network. The data can include data related to the health level of the ground material. The system can calculate the health level locally or using a remotely located device. The system can additionally transmit the calculated health level and other relevant data, such as the testing location, to a remotely located device using a wireless modem. The vehicle can include safety equipment. For example, safety equipmentcan include a light capable of flashing to signal the system's physical location. Additionally, safety equipmentcan include a sound system that generates auditory notifications such as beeping or a horn to notify nearby personnel about the location of the system or what the system is doing.

is an isometric view of an embodimentof the system in a first positionThe first positionis before augeris vertically translated toward ground material. In the first positiona motor within auger controllercan position augerso that the auger tipis positioned away from the ground material. For example, auger controllercan control the position of augerby rotating a thread through the motor. The system includes a distance sensor. Distance sensoris configured to detect a depth or distance from auger tipto ground material. Distance sensorcan include range sensors, a camera, a radar sensor, a laser distance sensor, an ultrasonic sensor, or any sensor capable of measuring a distance from one point to another. Distance sensoris also configured to detect the presence of any debristhat may be present in the path of auger. Distance sensorhas a detection window. Detection windowis the area where distance sensorcan detect the presence of debris.

is an isometric view of an embodimentof the system in a second positionThe second positionis after augerhas been vertically translated into the ground material. In the second positionthe motor within auger controllercan position augerso that the auger tipis positioned in the ground material. The distance sensorcan detect debrisin the detection window. Auger controllercan determine that the debrisis not in the path of auger. The system, therefore, determines that a new testing location is not needed, meaning the system can continue with the testing process. When auger controllerdetermines that debrisis in the path of auger, the auger controller can notify a user to adjust the testing location or cause the vehicle to move to a different testing location.

is a front-facing view of an embodiment of auger. Augercan include auger tip. Auger tipcan have a conical shape that comes to a point. The conical shape of auger tipallows augerto pierce the ground material with less resistance when compared to a flat or blunted tip. Auger tipcan be attached to or a part of auger shank. The auger shankcan have a cylindrical shape. In one embodiment, auger shankhas a uniform diameter until reaching auger tip. In another embodiment, auger shankhas a varying diameter. A varying diameter can allow augerto better accommodate sensoror better direct the ground material toward sensor window.

Auger bladeis attached to auger shank. Auger bladecan extend helically about the outer diameter of auger shank. For example, auger bladecan extend a predetermined distance along auger shankstarting from auger tip. In another example, auger bladecan be offset a predetermined distance from auger tipbefore extending along auger shank. Auger bladecan extend perpendicularly to or at a predetermined angle from the outer diameter of the auger shank. Auger bladecan have a different geometry based on the size of augeror the type of ground material at a location.

The auger bladecan include blade scoop. Blade scoopcan extend from the section of the auger blade in front of sensor window. Blade scoopcan extend vertically from the auger blade toward the sensor window. The geometry of blade scoopcan change based on the type of ground material and the size of augeror the sensor window. For example, blade scoopcan be a separate piece attached to the auger bladeor be a bent portion of auger blade. Blade scoopcan be configured to direct ground material toward sensor window. For example, blade scoopcan have a flat, concave, or convex geometry to direct different amounts of ground material toward sensor window. The geometry of blade scoopcan allow the ground material to pass over sensor windowwithout compacting the ground material against sensor window. Compacting the ground material can affect the reading of sensor, which is located behind sensor window, and cause ground material to build up in front of sensor window. A buildup of ground material between the sensor windowand blade scoopcan prevent multiple readings from being taken and can cause the need for manual cleaning of auger.

Sensor windowcan be constructed from a transparent material such as glass or plastic. The size of sensor windowcan be determined based on the type of sensorhoused behind the sensor windowor the size of auger. Sensorcan be connected to auger controllerusing wires. The wires can be configured to supply sensorwith electricity or facilitate the flow of data between the auger controller and sensor.

is a block diagram of an embodimentof the system. The system can receive GPS coordinates from a satelliteor another wireless network. The GPS coordinates correspond to the measurement points for a given area. The system can generate a navigation route. The navigation routecan be a series of driving instructions used to autonomously navigate the system using an off-road vehicleto the measurement area. For example, the navigation route can cause vehicleto drive on private property, known paths, public roads, or a combination of the three.

The system generates a measurement plan. The measurement plancan include all the measurement points for the measurement area. For example, measurement plancan include an expected or known topography of the measurement area. Vehiclecan, using the navigation route, drive to measurement area. The measurement areacan be cropland with or without crops or any area of land where a health level measurement needs to be taken. The system can begin taking measurements at the measurement area. The system can use a detection systemto determine if debris in the path of the auger could damage the system and prevent a measurement from being taken. The system can generate graphof the ground material health level. Graphcan be a scatterplot, a line graph, a bar chart, or a pie chart. Graphcan be generated for a single measurement at one measurement point, multiple measurements at one measurement point, a single measurement at multiple measurement points, or multiple measurements at multiple measurement points.

is a flowchart of an embodiment of the system. In one example, the system can be embodied in a computer system, the system including at least one hardware processor and at least one non-transitory memory storing instructions, which, when executed by the at least one hardware processor, cause the system to perform the process.

At step, the system can receive a request to measure a health level for ground material for an area. At step, the system can receive, from a global positioning system (GPS), coordinates for the area. At step, the system can determine a measurement plan. In one example, the measurement plan includes a location of at least one measurement point. At step, the system can execute the measurement plan. Executing the measurement plan causes the system to continue to the next step. At step, the system can navigate to the location of the at least one measurement point. At step, the system can detect for a presence of debris in a drilling path of an auger. The debris is a substance separate from the ground material. Drilling into the debris damages the auger. The system can adjust the location of the at least one measurement point to avoid the debris.

At step, the system can measure multiple characteristics of the ground material using the auger. The auger includes at least one sensor configured to measure the multiple characteristics of the ground material. In one example, the multiple characteristics of the ground material include a nutrient level, a chemical composition, a moisture level, a temperature, a compaction level, or a potential of hydrogen (pH) level of the ground material. In one example, the at least one sensor includes a spectrometer, a range sensor, a moisture sensor, a nutrient sensor, a force sensor, a camera, a temperature sensor, a potential of hydrogen (pH) sensor, or a chemical sensor.

At step, the system can calculate, using a machine-learning model, the health level of the ground material based on the measured characteristics of the ground material. In one example, the health level is calculated locally on the system. In one example, the system can receive control data for multiple ground material samples. The multiple ground material samples can have varying measurements for the multiple characteristics. The multiple ground material samples can have varying health levels. The system can train the machine-learning model based on the received control data for the multiple ground material samples for each of the multiple measurement points for the area. In another example, the system can generate a gradient map of the health level for the at least one measurement point at the area. The gradient map can indicate the health level of the ground material for multiple depths below a ground surface at the at least one measurement point.

In another example, the system can receive a request to measure a health level for the ground material at multiple areas. The system can determine a measurement plan for the multiple areas. The system can execute the measurement plan for the multiple areas. Executing the measurement plan for the multiple areas can further cause the system to measure the multiple characteristics of the ground material at each of the multiple areas using the auger. The system can calculate the health level of the ground material for each of the multiple areas. The system can rank the multiple areas based on the health level of the area. The system can generate, based on the ranking, a map of the multiple areas. The map can indicate which of the multiple areas is suitable for cultivating a plant.

In another example, the system can determine, based on the health level, an action plan for the ground material for the area. The action plan includes any one of the following: a recommendation to perform more measurements of the ground material at the area, a cultivation plan with recommendations of different plant species suitable to be cultivated at the area, or a ground material rejuvenation plan with a list of deficiencies in the ground material and a list of techniques to reduce the deficiencies in the ground material at the area. In another example, the system indicates the location to cultivate a plant and the type of plant to cultivate, thereby increasing the plant growth and thus reducing an amount of greenhouse gases by increasing the amount of greenhouse gases consumed by the plant.

In another example, the system can log the measurements of the multiple characteristics of the ground material and the health level for the at least one measurement point at the area. The system can perform a second measurement of the multiple characteristics of the ground material after a predetermined time period for the at least one measurement point for the area. The system can calculate a second health level of the ground material for the at least one measurement point at the area. The system can compare the logged measurements of the multiple characteristics and the health level of the ground material to the second measurement of the multiple characteristics and the second health level of the ground material. The system can generate a ground material report indicating a change in the health level of the ground material for the area over the predetermined time period.

is a block diagram illustrating an example ML system, in accordance with one or more embodiments. Likewise, different embodiments of the ML systeminclude different and/or additional components and are connected in different ways. The ML systemis sometimes referred to as an ML module.

The ML systemincludes a feature extraction moduleimplemented using components of the example computer systemillustrated and described in more detail with reference to. In some embodiments, the feature extraction moduleextracts a feature vectorfrom input data. The feature vectorincludes features. . . ,The feature extraction modulereduces the redundancy in the input data, for example, repetitive data values, to transform the input datainto the reduced set of features, for example, features. . . ,The feature vectorcontains the relevant information from the input datasuch that events or data value thresholds of interest are identified by the ML modelby using a reduced representation. In some example embodiments, the following dimensionality reduction techniques are used by the feature extraction module: independent component analysis, Isomap, kernel principal component analysis (PCA), latent semantic analysis, partial least squares, PCA, multifactor dimensionality reduction, nonlinear dimensionality reduction, multilinear PCA, multilinear subspace learning, semidefinite embedding, autoencoder, and deep feature synthesis.

In alternate embodiments, the ML modelperforms deep learning (also known as deep structured learning or hierarchical learning) directly on the input datato learn data representations, as opposed to using task-specific algorithms. In deep learning, no explicit feature extraction is performed; the featuresare implicitly extracted by the ML system. For example, the ML modeluses a cascade of multiple layers of nonlinear processing units for implicit feature extraction and transformation. Each successive layer uses the output from the previous layer as input. The ML modelthus learns in supervised (e.g., classification) and/or unsupervised (e.g., pattern analysis) modes. The ML modellearns multiple levels of representations that correspond to different levels of abstraction, wherein the different levels form a hierarchy of concepts. The multiple levels of representation configure the ML modelto differentiate features of interest from background features.

In alternative example embodiments, the ML model, for example, in the form of a CNN, generates the output, without the need for feature extraction, directly from the input data. The outputis provided to the computer device. The computer deviceis a server, computer, tablet, smartphone, smart speaker, etc., implemented using components of the example computer systemillustrated and described in more detail with reference to. In some embodiments, the steps performed by the ML systemare stored in memory on the computer devicefor execution. In other embodiments, the outputis displayed on an electronic display of the computer device.

A CNN is a type of feed-forward artificial neural network in which the connectivity pattern between its neurons is inspired by the organization of a visual cortex. Individual cortical neurons respond to stimuli in a restricted area of space known as the receptive field. The receptive fields of different neurons partially overlap such that they tile the visual field. The response of an individual neuron to stimuli within its receptive field is approximated mathematically by a convolution operation. CNNs are based on biological processes and are variations of multilayer perceptrons designed to use minimal amounts of preprocessing.

In embodiments, the ML modelis a CNN that includes both convolutional layers and max pooling layers. For example, the architecture of the ML modelis “fully convolutional,” which means that variable-sized sensor data vectors are fed into it. For convolutional layers, the ML modelspecifies a kernel size, a stride of the convolution, and an amount of zero padding applied to the input of that layer. For the pooling layers, the modelspecifies the kernel size and stride of the pooling.

In some embodiments, the ML systemtrains the ML model, based on the training data, to correlate the feature vectorto expected outputs in the training data. As part of the training of the ML model, the ML systemforms a training set of features and training labels by identifying a positive training set of features that have been determined to have a desired property in question and, in some embodiments, forms a negative training set of features that lack the property in question.

The ML systemapplies ML techniques to train the ML model, which, when applied to the feature vector, outputs indications of whether the feature vectorhas an associated desired property or properties, such as a probability that the feature vectorhas a particular Boolean property, or an estimated value of a scalar property. In embodiments, the ML systemfurther applies dimensionality reduction (e.g., via linear discriminant analysis (LDA), PCA, or the like) to reduce the amount of data in the feature vectorto a smaller, more representative set of data.

In embodiments, the ML systemuses supervised ML to train the ML model, with feature vectors of the positive training set and the negative training set serving as the inputs. In some embodiments, different ML techniques, such as linear support vector machine (linear SVM), boosting for other algorithms (e.g., AdaBoost), logistic regression, naïve Bayes, memory-based learning, random forests, bagged trees, decision trees, boosted trees, boosted stumps, neural networks, CNNs, etc., are used. In some example embodiments, a validation setis formed of additional features other than those in the training data, which have already been determined to have or to lack the property in question. The ML systemapplies the trained ML modelto the features of the validation setto quantify the accuracy of the ML model. Common metrics applied in accuracy measurement include Precision and Recall, where Precision refers to a number of results the ML modelcorrectly predicted out of the total it predicted and Recall is a number of results the ML modelcorrectly predicted out of the total number of features that had the desired property in question. In some embodiments, the ML systemiteratively retrains the ML modeluntil the occurrence of a stopping condition, such as the accuracy measurement indication that the ML modelis sufficiently accurate, or a number of training rounds have taken place. In embodiments, the validation setincludes data corresponding to confirmed locations, dates, times, activities, or combinations thereof. This allows the detected values to be validated using the validation set. The validation setis generated based on the analysis to be performed.

is a block diagram that illustrates an example of a computer systemin which at least some operations described herein can be implemented. As shown, the computer systemcan include: one or more processors, main memory, non-volatile memory, a network interface device, a video display device, an input/output device, a control device(e.g., keyboard and pointing device), a drive unitthat includes a machine-readable (storage) medium, and a signal generation devicethat are communicatively connected to a bus. The busrepresents one or more physical buses and/or point-to-point connections that are connected by appropriate bridges, adapters, or controllers. Various common components (e.g., cache memory) are omitted fromfor brevity. Instead, the computer systemis intended to illustrate a hardware device on which components illustrated or described relative to the examples of the figures and any other components described in this specification can be implemented.

The computer systemcan take any suitable physical form. For example, the computing systemcan share a similar architecture as that of a server computer, personal computer (PC), tablet computer, mobile telephone, game console, music player, wearable electronic device, network-connected (“smart”) device (e.g., a television or home assistant device), AR/VR systems (e.g., head-mounted display), or any electronic device capable of executing a set of instructions that specify action(s) to be taken by the computing system. In some implementations, the computer systemcan be an embedded computer system, a system-on-chip (SOC), a single-board computer system (SBC), or a distributed system such as a mesh of computer systems, or it can include one or more cloud components in one or more networks. Where appropriate, one or more computer systemscan perform operations in real time, in near real time, or in batch mode.

The network interface deviceenables the computing systemto mediate data in a networkwith an entity that is external to the computing systemthrough any communication protocol supported by the computing systemand the external entity. Examples of the network interface deviceinclude a network adapter card, a wireless network interface card, a router, an access point, a wireless router, a switch, a multilayer switch, a protocol converter, a gateway, a bridge, a bridge router, a hub, a digital media receiver, and/or a repeater, as well as all wireless elements noted herein.

The memory (e.g., main memory, non-volatile memory, machine-readable medium) can be local, remote, or distributed. Although shown as a single medium, the machine-readable mediumcan include multiple media (e.g., a centralized/distributed database and/or associated caches and servers) that store one or more sets of instructions. The machine-readable mediumcan include any medium that is capable of storing, encoding, or carrying a set of instructions for execution by the computing system. The machine-readable mediumcan be non-transitory or comprise a non-transitory device. In this context, a non-transitory storage medium can include a device that is tangible, meaning that the device has a concrete physical form, although the device can change its physical state. Thus, for example, non-transitory refers to a device remaining tangible despite this change in state.

Although implementations have been described in the context of fully functioning computing devices, the various examples are capable of being distributed as a program product in a variety of forms. Examples of machine-readable storage media, machine-readable media, or computer-readable media include recordable-type media such as volatile and non-volatile memory, removable flash memory, hard disk drives, optical disks, and transmission-type media such as digital and analog communication links.

In general, the routines executed to implement examples herein can be implemented as part of an operating system or a specific application, component, program, object, module, or sequence of instructions (collectively referred to as “computer programs”). The computer programs typically comprise one or more instructions (e.g., instructions,,) set at various times in various memory and storage devices in computing device(s). When read and executed by the processor, the instruction(s) cause the computing systemto perform operations to execute elements involving the various aspects of the disclosure.

The terms “example,” “embodiment,” and “implementation” are used interchangeably. For example, references to “one example” or “an example” in the disclosure can be, but not necessarily are, references to the same implementation; and such references mean at least one of the implementations. The appearances of the phrase “in one example” are not necessarily all referring to the same example, nor are separate or alternative examples mutually exclusive of other examples. A feature, structure, or characteristic described in connection with an example can be included in another example of the disclosure. Moreover, various features are described that can be exhibited by some examples and not by others. Similarly, various requirements are described that can be requirements for some examples but not for other examples.